PROJECT SUMMARY Depression is a heterogeneous condition with varying degrees of severity and different etiologies that contrib- ute to variability in course trajectory and psychosocial functioning. About a third of adolescents with depression exhibit manic symptoms and as such, this substantially increases the risk for psychosocial impairment and sui- cidality. Yet, few studies have examined the pathophysiology of manic symptoms in adolescent depression and how alterations in key neural systems may be associated with changes in emotional, cognitive, and behavioral functioning. This project will address this important gap by employing a dimensional approach and a longitudi- nal design to examine neural, emotional, cognitive, and behavioral substrates of manic symptoms in (non- bipolar) depressed adolescents. Given evidence that manic symptoms implicate dysregulation of emotion and behavior, the current project will test the overarching hypothesis that alterations in connectivity of fronto- striatal-limbic (FSL) circuitry underlying cognitive control of positive emotion and reward and associated altera- tions in RDoC domain-related (emotional, cognitive, behavioral) measures of functioning in Cognitive and Posi- tive Valence Systems could differentiate depressed adolescents with from those without co-occurring manic symptoms. We aim to test this model in a longitudinal study of 180 mid-/post-pubertal adolescents (12-17 years old), including 140 adolescents with clinically significant depressive symptoms varying in manic symp- toms and 40 age- and sex- matched healthy youth. In accordance with the NIMH RDoC initiative, all depressed adolescents will be recruited from treatment settings based on the presence of clinically significant depression symptoms, irrespective of their categorical depressive disorder diagnosis. The project includes at baseline: a) fMRI paradigms measuring functioning of FSL circuitry associated with cognitive control of positive emotion and reward, b) resting state and diffusion tensor imaging, and c) out-of-the scanner neurocognitive tasks (e.g., approach-avoidance). RDoC domain-related emotional (e.g., elated mood), cognitive (e.g., delay discounting), and behavioral (e.g., impulsivity) functioning and levels of arousal will be assessed using a multi-method ap- proach (e.g., computer tasks in the lab and ecologically-valid web-based mood and actigraphy measures) at baseline and repeated bi-annually over 2 years post-scan. In addition, using machine learning algorithms, ex- ploratory analyses will be performed to identify patterns of baseline neuroimaging data and trajectories of RDoC domain-related (emotional, cognitive, behavioral) measures that can predict levels of mood symptoms and functional impairment at 2 years. Findings from this study could help distinguish subgroups of depressed adolescents based on brain-behavioral relationships, improve early detection of manic symptoms in depressed youth, and inform personalized interventions (e.g., antidepressant vs. mood-stabilizer, type of psychotherapy).